Hi all,
2015-10-30 20:59 GMT+03:00 Matthew Hall :
> On Fri, Oct 30, 2015 at 12:00:18PM +, Bruce Richardson wrote:
> > Matthew's patches were attachments, I don't think they came through in
> patchwork
> > correctly :-(, but that is the relevant link there anyway.]
>
> Let me know if there is
To support FVL PMD can select which RX and TX function should be used
according to the queue config.
Signed-off-by: Zhe Tao
---
doc/guides/rel_notes/release_2_2.rst | 4 +
drivers/net/i40e/i40e_ethdev.c | 20 +++-
drivers/net/i40e/i40e_ethdev.h | 6 ++
drivers/net/i40e/i40e_ethd
To support the multiple segments in one packets when the received pkts exceed
one buffer size.
Signed-off-by: Zhe Tao
---
drivers/net/i40e/i40e_rxtx.c | 9 +++
drivers/net/i40e/i40e_rxtx.h | 3 +
drivers/net/i40e/i40e_rxtx_vec.c | 118 ---
3 files
The way to increase the performance of the vPMD TX is to use some fast mbuf
release method compares to the scalar TX.
Signed-off-by: Zhe Tao
---
drivers/net/i40e/i40e_rxtx.c | 8 ++
drivers/net/i40e/i40e_rxtx.h | 3 +
drivers/net/i40e/i40e_rxtx_vec.c | 163 ++
The vPMD RX function uses the multi-buffer and SSE instructions to
accelerate the RX speed, but now the pktype cannot be supported by the vPMD RX,
because it will decrease the performance heavily.
Signed-off-by: Zhe Tao
---
config/common_bsdapp | 2 +
config/common_linuxapp
This patch set add the vector PMD support for FVL.
FVL vPMD works like the way ixgbe does
All the functionality is tested
PATCH v1: add the vector PMD support for FVL
PATCH v2: remove the extra code out of share code
change the prefetch position for splitter packets support
u
On Fri, Oct 30, 2015 at 10:34:54PM +0300, Vladimir Medvedkin wrote:
> Hi all,
>
> 2015-10-30 20:59 GMT+03:00 Matthew Hall :
>
> > On Fri, Oct 30, 2015 at 12:00:18PM +, Bruce Richardson wrote:
> > > Matthew's patches were attachments, I don't think they came through in
> > patchwork
> > > corr
Fixes an initialization issue of 'valid_group' in the delete_depth_small
function.
In this function, use new rte_lpm_tbl8_entry we call A to replace the old
rte_lpm_tbl8_entry. But the valid_group do not set VALID, so it
will be INVALID.
Then when adding a new route which depth is > 24,the tbl
Fixes an issue of check logic in delete_depth_small function.
For a tbl24 entry, the 'ext_entry' field indicates whether we need to use
tbl8_gindex to read the next_hop from a tbl8 entry, or whether it can be read
directly from this entry.
If a route is deleted, the prefix of previous route is
Fixes two issues in the delete_depth_small() function.
v2 changes:
Split a patch into two patches for two issues.
Add more clear issue description.
*** BLURB HERE ***
Jijiang Liu (2):
fix an issue of condition check in delete_depth_small().
fix an initialization issue of valid_group in
To support FVL PMD can select which RX and TX function should be used
according to the queue config.
Signed-off-by: Zhe Tao
---
doc/guides/rel_notes/release_2_2.rst | 4 +
drivers/net/i40e/i40e_ethdev.c | 20 +++-
drivers/net/i40e/i40e_ethdev.h | 6 ++
drivers/net/i40e/i40e_ethd
To support the multiple segments in one packets when the received pkts exceed
one buffer size.
Signed-off-by: Zhe Tao
---
drivers/net/i40e/i40e_rxtx.c | 9
drivers/net/i40e/i40e_rxtx.h | 3 ++
drivers/net/i40e/i40e_rxtx_vec.c | 103 +++
3 fil
The way to increase the performance of the vPMD TX is to use some fast mbuf
release method compares to the scalar TX.
Signed-off-by: Zhe Tao
---
drivers/net/i40e/i40e_rxtx.c | 8 ++
drivers/net/i40e/i40e_rxtx.h | 3 +
drivers/net/i40e/i40e_rxtx_vec.c | 162 ++
The vPMD RX function uses the multi-buffer and SSE instructions to
accelerate the RX speed, but now the pktype cannot be supported by the vPMD RX,
because it will decrease the performance heavily.
Signed-off-by: Zhe Tao
---
config/common_bsdapp | 2 +
config/common_linuxapp
This patch set add the vector PMD support for FVL.
FVL vPMD works like the way ixgbe does
All the functionality is tested
PATCH v1: add the vector PMD support for FVL
PATCH v2: remove the extra code out of share code
change the prefetch position for splitter packets support
u
Modify MAC address of csum forwarding.
The change will affect on the csum fwd performance.
But I also think the change is necessary, or we cannot use csumonly fwd mode in
a VM.
Signed-off-by: Jijiang Liu
---
app/test-pmd/csumonly.c |6 ++
1 files changed, 6 insertions(+), 0 deletions(-
Change the vhost sample to support and test TX offload.
Signed-off-by: Jijiang Liu
---
examples/vhost/main.c | 128 ++---
1 files changed, 111 insertions(+), 17 deletions(-)
diff --git a/examples/vhost/main.c b/examples/vhost/main.c
index 9eac2d0..06
Dequeue vhost TX offload in vhost lib.
Signed-off-by: Jijiang Liu
---
lib/librte_vhost/vhost_rxtx.c | 108 -
1 files changed, 107 insertions(+), 1 deletions(-)
diff --git a/lib/librte_vhost/vhost_rxtx.c b/lib/librte_vhost/vhost_rxtx.c
index 7026bfa..a888
Enqueue vhost TX checksum and TSO4/6 offload in virtio-net lib.
Signed-off-by: Jijiang Liu
---
drivers/net/virtio/virtio_rxtx.c | 61 ++
1 files changed, 61 insertions(+), 0 deletions(-)
diff --git a/drivers/net/virtio/virtio_rxtx.c b/drivers/net/virtio/vir
Add vhost TSO support capability in vhost lib.
Signed-off-by: Jijiang Liu
---
lib/librte_vhost/virtio-net.c |6 +-
1 files changed, 5 insertions(+), 1 deletions(-)
diff --git a/lib/librte_vhost/virtio-net.c b/lib/librte_vhost/virtio-net.c
index d0f1764..a5989a2 100644
--- a/lib/librte_
Fill virtio device info for TX offload.
Signed-off-by: Jijiang Liu
---
drivers/net/virtio/virtio_ethdev.c | 10 ++
1 files changed, 10 insertions(+), 0 deletions(-)
diff --git a/drivers/net/virtio/virtio_ethdev.c
b/drivers/net/virtio/virtio_ethdev.c
index cb5dfee..b831c02 100644
---
Add vhost TX checksum and TSO capabilities in virtio-net lib.
Signed-off-by: Jijiang Liu
---
drivers/net/virtio/virtio_ethdev.h |5 -
1 files changed, 4 insertions(+), 1 deletions(-)
diff --git a/drivers/net/virtio/virtio_ethdev.h
b/drivers/net/virtio/virtio_ethdev.h
index 9026d42..6ee
Record virtual address of virtio net header.
Signed-off-by: Jijiang Liu
---
drivers/net/virtio/virtio_ethdev.c |3 +++
1 files changed, 3 insertions(+), 0 deletions(-)
diff --git a/drivers/net/virtio/virtio_ethdev.c
b/drivers/net/virtio/virtio_ethdev.c
index 465d3cd..cb5dfee 100644
--- a/d
The virtual addr for virtio net header need to be recorded.
Signed-off-by: Jijiang Liu
---
drivers/net/virtio/virtqueue.h |1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/drivers/net/virtio/virtqueue.h b/drivers/net/virtio/virtqueue.h
index 7789411..530f840 100644
--- a/d
Adds vhost TX offload support.
The patch set add the negotiation between us-vhost and virtio-net for vhost TX
offload(checksum and TSO), and add the TX offload support in the libs and
change vhost sample and csum application to test these changes.
v2 change:
fill virtio device information for
2015-10-29 18:33, Xie, Huawei:
> On 10/29/2015 2:34 AM, Pavel Boldin wrote:
> > The patchset contains an attempt at the refactoring the `eventfd_link`
> > kernel module that is used to steal an FD in DPDK.
> >
> > The first patch refactors old EVENTFD_COPY handler fixing the codepath
> > and errors
Fixes: 7fae69eeff13 ("mlx4: new poll mode driver")
Signed-off-by: Adrien Mazarguil
---
drivers/net/mlx4/mlx4.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/net/mlx4/mlx4.c b/drivers/net/mlx4/mlx4.c
index 3b25492..691e0e5 100644
--- a/drivers/net/mlx4/mlx4.c
+++ b/drivers/net/mlx4/
Port is considering inactive when the related netdevice is down. There is no
reason to warn about it (and confuse users) since it's automatically brought
up later by the PMD.
Signed-off-by: Adrien Mazarguil
---
drivers/net/mlx4/mlx4.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
From: Nelio Laranjeiro
Prefetching initial bytes of mbuf structures earlier and in two cache lines
instead of one improves performance of mlx4_rx_burst(), which accesses the
mbuf->next field not present in the first 128 bytes.
Signed-off-by: Nelio Laranjeiro
---
drivers/net/mlx4/mlx4.c | 11 ++
From: Nelio Laranjeiro
ConnectX-4 is able to use indirection tables size of power of two, but
with the current API it is impossible to predict its size, so to simplify,
for any query/update RETA command, the indirection table is modified to use
the maximum value.
A port stop/start must be done t
From: Nelio Laranjeiro
Several NICs can handle 512 entries in their RETA table, an 8 bit field is
not large enough for them.
Signed-off-by: Nelio Laranjeiro
---
lib/librte_ether/rte_ethdev.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/librte_ether/rte_ethdev.h b/lib
From: Nelio Laranjeiro
For RETA query/update with a table of 512 entries, buffers are too small to
handle the request.
Signed-off-by: Nelio Laranjeiro
---
lib/librte_cmdline/cmdline_parse.h| 2 +-
lib/librte_cmdline/cmdline_parse_string.h | 2 +-
lib/librte_cmdline/cmdline_rdline.h
mlx5 devices support indirection tables of variable size up to 512 entries,
which requires a larger configuration structure (requiring a change in the
ABI).
This patchset can be considered as a first RFC step because the current API
is not very practical due to the following limitations:
- Config
From: Nelio Laranjeiro
Add interrupts handler for port status notification.
Signed-off-by: Nelio Laranjeiro
Signed-off-by: Adrien Mazarguil
---
drivers/net/mlx5/mlx5.c | 3 +
drivers/net/mlx5/mlx5.h | 7 ++
drivers/net/mlx5/mlx5_defs.h| 3 +
drivers/net/mlx5/mlx5_eth
From: Nelio Laranjeiro
Add interrupts handler for port status notification.
Signed-off-by: Nelio Laranjeiro
---
drivers/net/mlx4/mlx4.c | 166
drivers/net/mlx4/mlx4.h | 3 +
2 files changed, 169 insertions(+)
diff --git a/drivers/net/mlx4/mlx
From: Nelio Laranjeiro
Add RTE_INTR_HANDLE_EXT handler type for PMDs that do not support VFIO or
UIO. Those are expected to manage the file descriptor themselves.
Signed-off-by: Nelio Laranjeiro
---
lib/librte_eal/linuxapp/eal/eal_interrupts.c | 34 --
.../linuxapp/e
Mellanox PMDs (mlx4 and mlx5) support link status notification with
libibverbs through a file descriptor, in a manner reminiscent of UIO/VFIO.
It still requires a new "external" interrupt handler type since contrary
to other types, this file descriptor must not be processed by EAL.
Changes in v2:
Add new features related to Mellanox OFED 3.1 support.
Signed-off-by: Adrien Mazarguil
---
doc/guides/nics/mlx5.rst | 26 ++
1 file changed, 26 insertions(+)
diff --git a/doc/guides/nics/mlx5.rst b/doc/guides/nics/mlx5.rst
index fdb621c..2d68914 100644
--- a/doc/guides/n
From: Yaacov Hazan
Seen with GCC < 4.6:
error: unknown field ?tcp_udp? specified in initializer
error: extra brace group at end of initializer
Static initialization of anonymous structs/unions is a C11 feature
properly supported only since GCC 4.6.
Work around compilation errors with older v
For adapters that support it, this flag improves performance outside of VF
context.
Signed-off-by: Adrien Mazarguil
---
drivers/net/mlx5/Makefile | 5 +
drivers/net/mlx5/mlx5_txq.c | 7 +++
2 files changed, 12 insertions(+)
diff --git a/drivers/net/mlx5/Makefile b/drivers/net/mlx5/Mak
Normal flows do not currently provide IPv6 support.
Signed-off-by: Adrien Mazarguil
Signed-off-by: Nelio Laranjeiro
---
drivers/net/mlx5/Makefile | 4 +++
drivers/net/mlx5/mlx5_mac.c| 18 ++--
drivers/net/mlx5/mlx5_rxmode.c | 22 +++
drivers/net/mlx5/mlx5_rxq.c
From: Nelio Laranjeiro
Only a single flow per hash RX queue is needed in promiscuous mode.
Disable others to free up hardware resources.
Signed-off-by: Nelio Laranjeiro
Signed-off-by: Adrien Mazarguil
Signed-off-by: Yaacov Hazan
---
drivers/net/mlx5/mlx5_mac.c | 5 +
drivers/net/mlx
The "show port X rss-hash" command sometimes displays garbage instead of the
expected RSS hash key because the maximum key length is undefined. When the
requested key is too large to fit in the buffer,
rte_eth_dev_rss_hash_conf_get() does not update it.
Signed-off-by: Adrien Mazarguil
Signed-off-
From: Nelio Laranjeiro
DPDK uses a structure to get or set a new hash key (see
eth_rte_rss_hash_conf). rss_hf field from this structure is used in
rss_hash_get_conf to retrieve the hash key and in rss_hash_update uses
it to verify the key exists before trying to update it.
Signed-off-by: Nelio
From: Nelio Laranjeiro
DPDK expects to have an RSS hash key per flow type (IPv4, IPv6, UDPv4,
etc.), to handle this the PMD must keep a table of hash keys to be able
to reconfigure the queues at each start/stop call.
Signed-off-by: Nelio Laranjeiro
Signed-off-by: Adrien Mazarguil
---
drivers/
From: Nelio Laranjeiro
First implementation of rss_hash_update and rss_hash_conf_get, those
functions still lack in functionality but are usable to change the RSS
hash key. For now, the PMD does not handle an indirection table for
each kind of flow (IPv4, IPv6, etc.), the same RSS hash key is us
From: Olga Shern
Promiscuous and allmulticast modes were historically enabled by adding
specific flows with types IBV_FLOW_ATTR_ALL_DEFAULT or
IBV_EXP_FLOW_ATTR_MC_DEFAULT to each hash RX queue, but this method is
deprecated.
- Promiscuous mode is now enabled by omitting destination MAC addresse
From: Olga Shern
All hash RX QPs currently use the same flow steering rule (L2 MAC filtering)
regardless of their type (TCP, UDP, IPv4, IPv6), which prevents them from
being dispatched properly. This is fixed by adding flow information to the
hash RX queue initialization data and generating speci
From: Nelio Laranjeiro
Use the maximum size of the indirection table when the number of requested
RX queues is not a power of two, this help to improve RSS balancing.
A message informs users that balancing is not optimal in such cases.
Signed-off-by: Nelio Laranjeiro
Signed-off-by: Adrien Maza
From: Olga Shern
The default hash RX queue handles packets that are not matched by more
specific types and requires its own indirection table of size 1 to work
properly.
This commit implements support for multiple indirection tables by grouping
their layout and properties in a static initializat
The new Verbs RSS API is lower-level than the previous one and much more
flexible but requires RX queues to use Work Queues (WQs) internally instead
of Queue Pairs (QPs), which are grouped in an indirection table used by a
new kind of hash RX QPs.
Hash RX QPs and the indirection table together rep
Removing this structure reduces the size of SG and non-SG RX queue elements
significantly to improve performance.
An nice side effect is that the mbuf pointer is now fully stored in
struct rxq_elt instead of relying on the WR ID data offset hack.
Signed-off-by: Adrien Mazarguil
Signed-off-by: Ol
This commit updates mlx5_rx_burst_sp() to use the fast verbs interface for
posting RX buffers just like mlx5_rx_burst(). Doing so avoids a loop in
libmlx5 and an indirect function call through libibverbs.
Note: recv_sg_list() is not implemented in the QP burst API, this commit is
only to prepare t
Mellanox OFED 3.1 [1] comes with improved APIs that Mellanox ConnectX-4
(mlx5) adapters can take advantage of, such as:
- Separate post and doorbell operations on all queues.
- Lightweight RX queues called Work Queues (WQs).
- Low-level RSS indirection table and hash key configuration.
This patch
Signed-off-by: Adrien Mazarguil
---
doc/guides/nics/mlx5.rst | 308 +++
doc/guides/rel_notes/release_2_2.rst | 9 +
2 files changed, 317 insertions(+)
create mode 100644 doc/guides/nics/mlx5.rst
diff --git a/doc/guides/nics/mlx5.rst b/doc/guides/nic
This is the same implementation as mlx4.
Signed-off-by: Adrien Mazarguil
Signed-off-by: Nelio Laranjeiro
---
drivers/net/mlx5/mlx5_rxq.c | 14 +++
drivers/net/mlx5/mlx5_rxtx.c | 94 +++
drivers/net/mlx5/mlx5_rxtx.h | 2 +
drivers/net/mlx5/mlx5_ut
All MAC RX flows must be updated with VLAN information when configuring a
VLAN filter.
Signed-off-by: Adrien Mazarguil
Signed-off-by: Nelio Laranjeiro
---
drivers/net/mlx5/Makefile| 1 +
drivers/net/mlx5/mlx5.c | 1 +
drivers/net/mlx5/mlx5.h | 6 ++
drivers/net/mlx5/mlx5_def
Like most other device control operations, those are handled by the related
kernel network device through syscalls.
Signed-off-by: Adrien Mazarguil
Signed-off-by: Nelio Laranjeiro
---
drivers/net/mlx5/mlx5.c| 2 +
drivers/net/mlx5/mlx5.h| 2 +
drivers/net/mlx5/mlx5_ethdev.c |
Link information is retrieved using ethtool ioctls.
Signed-off-by: Adrien Mazarguil
Signed-off-by: Nelio Laranjeiro
---
drivers/net/mlx5/mlx5.c| 1 +
drivers/net/mlx5/mlx5.h| 1 +
drivers/net/mlx5/mlx5_ethdev.c | 71 ++
3 files changed,
These modes require special non-MAC flows.
Signed-off-by: Adrien Mazarguil
Signed-off-by: Nelio Laranjeiro
Signed-off-by: Yaacov Hazan
---
drivers/net/mlx5/Makefile | 1 +
drivers/net/mlx5/mlx5.c | 4 +
drivers/net/mlx5/mlx5.h | 13 ++
drivers/net/mlx5/mlx5_ethdev.c
Hardware counters are not supported yet.
Signed-off-by: Adrien Mazarguil
Signed-off-by: Nelio Laranjeiro
---
drivers/net/mlx5/Makefile | 1 +
drivers/net/mlx5/mlx5.c | 2 +
drivers/net/mlx5/mlx5.h | 5 ++
drivers/net/mlx5/mlx5_defs.h | 8 +++
drivers/net/mlx5/mlx5_rxq.c
Depending on the MTU and whether jumbo frames are enabled, RX queues may
switch between SG and non-SG modes for better performance.
Signed-off-by: Adrien Mazarguil
Signed-off-by: Nelio Laranjeiro
---
drivers/net/mlx5/mlx5.c| 1 +
drivers/net/mlx5/mlx5.h| 1 +
drivers/net/mlx
A dedicated RX callback is added to handle scattered buffers. For better
performance, it is only used when jumbo frames are enabled and MTU is larger
than a single mbuf.
On the TX path, scattered buffers are also handled in a separate function.
When there are more than MLX5_PMD_SGE_WR_N segments i
This commit adds the remaining missing callbacks to make mlx5 usable.
Like mlx4, device start and stop are implemented on top of MAC RX flows.
Signed-off-by: Adrien Mazarguil
Signed-off-by: Nelio Laranjeiro
Signed-off-by: Francesco Santoro
Signed-off-by: Didier Pallard
---
drivers/net/mlx5/Ma
This commit adds support for MAC flow steering rules mandatory for the RX
path as well as the related callbacks to add/remove MAC addresses.
Signed-off-by: Adrien Mazarguil
Signed-off-by: Nelio Laranjeiro
Signed-off-by: Didier Pallard
---
drivers/net/mlx5/mlx5.c | 4 +-
drivers/net/mlx5
RSS implementation with parent/child QPs comes from mlx4 and is temporary.
Signed-off-by: Adrien Mazarguil
Signed-off-by: Nelio Laranjeiro
---
config/common_bsdapp | 3 +
config/common_linuxapp| 3 +
drivers/net/mlx5/Makefile | 15 +
drivers/net/mlx5/mlx5.c | 40
In its current state, this driver implements the bare minimum to initialize
itself and Mellanox ConnectX-4 adapters without doing anything else
(no RX/TX for instance). It is disabled by default since it is based on the
mlx4 driver and also depends on libibverbs.
Signed-off-by: Adrien Mazarguil
S
This PMD adds basic support for Mellanox ConnectX-4 (mlx5) families of
10/25/40/50/100 Gb/s adapters through the Verbs framework.
Its design is very similar to that of mlx4 from which most of its code is
borrowed without the mistake of putting it all in a single huge file.
It is disabled by defau
2015-10-27 16:54, Tetsuya Mukawa:
> Below patch has been submitted as a separate patch.
>
> - [dpdk-dev,1/3] vhost: Fix wrong handling of virtqueue array index
> (http://dpdk.org/dev/patchwork/patch/8038/)
Please could you rebase only the two last patches?
Thanks
PS:
WARNING:TYPO_SPELLING:
Sorry, there is a clang error.
2015-10-29 22:53, Huawei Xie:
> + _mm_prefetch((const void *)rused, _MM_HINT_T0);
virtio_rxtx_simple.c:197:2: error: cast from 'const void *' to
'void *' drops const qualifier
2015-10-29 22:53, Huawei Xie:
> +static int use_simple_rxtx;
error: ?use_simple_rxtx? defined but not used
I'll try to move it in the next patch.
Sorry there is a gcc error.
2015-10-18 22:16, Stephen Hemminger:
> + for (i = 0; i < vq_size; i++) {
virtio_ethdev.c:386:17: error: comparison between signed and unsigned integer
expressions
I see that the Hyper-V driver is on the roadmap for 2.2. I do not see any code
yet in the git development branch or patchwork. So two questions:
1. Is it likely we are going to see Hyper-V support in 2.2?
2. If we do see Hyper-V support in 2.2 will it support Rx Interrupt?
Regards
John
update the release note
Signed-off-by: Zhe Tao
---
doc/guides/rel_notes/release_2_2.rst | 4
1 file changed, 4 insertions(+)
diff --git a/doc/guides/rel_notes/release_2_2.rst
b/doc/guides/rel_notes/release_2_2.rst
index bc9b00f..c78ceca 100644
--- a/doc/guides/rel_notes/release_2_2.rst
++
move all the exta definition out of share code
Signed-off-by: Zhe Tao
---
drivers/net/i40e/base/i40e_type.h | 13 -
drivers/net/i40e/i40e_ethdev.c| 11 ++-
drivers/net/i40e/i40e_ethdev.h| 6 ++
drivers/net/i40e/i40e_ethdev_vf.c | 11 ++-
drivers/net/i40e/
according to ixgbe vector implementation, change the data prefetch position
Signed-off-by: Zhe Tao
---
drivers/net/i40e/i40e_rxtx_vec.c | 14 +++---
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/drivers/net/i40e/i40e_rxtx_vec.c b/drivers/net/i40e/i40e_rxtx_vec.c
index 6f8
edit the comments for the minimum packets support for vector RX
Signed-off-by: Zhe Tao
---
drivers/net/i40e/i40e_rxtx_vec.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/i40e/i40e_rxtx_vec.c b/drivers/net/i40e/i40e_rxtx_vec.c
index b2993e9..6f89002 100644
--- a/
To support FVL PMD can select which RX and TX function should be used
according to the queue config.
Signed-off-by: Zhe Tao
---
drivers/net/i40e/base/i40e_type.h | 10 +++
drivers/net/i40e/i40e_ethdev.c| 19 +++-
drivers/net/i40e/i40e_ethdev_vf.c | 27 --
drivers/net/i40e/i40e_rxtx.c
To support the multiple segments in one packets when the received pkts exceed
one buffer size.
Signed-off-by: Zhe Tao
---
drivers/net/i40e/i40e_rxtx.c | 9
drivers/net/i40e/i40e_rxtx.h | 3 ++
drivers/net/i40e/i40e_rxtx_vec.c | 103 +++
3 fil
The way to increase the performance of the vPMD TX is to use some fast mbuf
release method compares to the scalar TX.
Signed-off-by: Zhe Tao
---
drivers/net/i40e/i40e_rxtx.c | 8 ++
drivers/net/i40e/i40e_rxtx.h | 3 +
drivers/net/i40e/i40e_rxtx_vec.c | 162 ++
The vPMD RX function uses the multi-buffer and SSE instructions to
accelerate the RX speed, but now the pktype cannot be supported by the vPMD RX,
because it will decrease the performance heavily.
Signed-off-by: Zhe Tao
---
config/common_bsdapp | 2 +
config/common_linuxapp
This patch set add the vector PMD support for FVL.
FVL vPMD works like the way ixgbe does
All the functionality is tested
Zhe Tao (8):
add vector PMD RX for FVL
add vector PMD TX for FVL
add vector PMD scatter RX for FVL
add RX and TX selection function for FVL
edit the comments
chang
2015-10-18 10:04, Michael S. Tsirkin:
> On Fri, Oct 16, 2015 at 02:52:30PM +0100, Bruce Richardson wrote:
> > On Thu, Oct 15, 2015 at 04:18:59PM +0300, Michael S. Tsirkin wrote:
> > > On Thu, Oct 15, 2015 at 02:08:39PM +0300, Marcel Apfelbaum wrote:
> > > > Make vhost-user virtio 1.0 compatible by
2015-10-29 11:45, John Daley:
> For every packet sent, a completion was being requested and the
> posted_index register on the nic was being updated. Instead, request a
> completion and update the posted index once per burst after all
> packets have been sent by the burst function.
>
> Acked-by: S
2015-10-30 17:07, Iremonger, Bernard:
> Hi Thomas,
>
> > -Original Message-
> > From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> > Sent: Friday, October 30, 2015 5:01 PM
> > To: Iremonger, Bernard
> > Cc: dev at dpdk.org
> > Subject: Re: [dpdk-dev] [PATCH v2 1/1] vhost_xen: f
I have been attempting to build testpmd with the bnx2x pmd enabled to try and
talk through a bcm57810 10G NIC card. I have tried using SR-IOV VFs and I have
tried using the PFs. Neither seem to be able to be properly initialized in
testpmd and/or testpmd aborts.
Is this PMD actually usable or i
2015-10-22 14:21, Bernard Iremonger:
> changes in v2:
> update release notes
>
> Signed-off-by: Bernard Iremonger
Please what is the error?
> - rte_memcpy((void *)(uintptr_t)buff_hdr_addr, (const
> void*)&virtio_hdr, vq->vhost_hlen);
> + rte_memcpy((void *)(uintptr_t)bu
Hi dave
My name is Fares Bagh. I am from Freescale networking division.
We are very interested and supportive in the proposal below.
Our main interest is enabling HW acceleration options for our customers
starting with crypto function. We like to have a road map for acceleration
beyond crypto.
We
2015-10-30 05:42, Xie, Huawei:
> On 10/28/2015 12:07 AM, Bernard Iremonger wrote:
> > add Port Hotplug support to the xenvirt PMD
> >
> > This patch depends on v5 of the following patch set:
> >
> > remove-pci-driver-from-vdevs.patch
> >
> > Changes in v2:
> > Rebase
> > Update release notes.
> Ac
>
> These variables are needed to be able to manage one of virtio devices
> using both vhost library APIs and vhost PMD.
> For example, if vhost PMD uses current callback handler and private data
> provided by vhost library, A DPDK application that links vhost library
> cannot use some of vhost li
2015-10-30 16:16, Bruce Richardson:
> Another request, can you perhaps also fix the script for situations where
> checkpatch.pl is not in the kernel tree. I've used this script now to check a
> couple of patchsets, which came back clean, but it turns out that because I
> was
> using checkpatch.pl
Update interrupt related limitations for e1000em nic.
It only support one interrupt source and not support auto-clear.
Signed-off-by: Marvin Liu
diff --git a/doc/guides/nics/e1000em.rst b/doc/guides/nics/e1000em.rst
index fedafad..a84e8e1 100644
--- a/doc/guides/nics/e1000em.rst
+++ b/doc/guides
Sorry it does not compile with clang.
2015-10-27 17:21, Shaopeng He:
> + for (i = 0; i < (int)vmdq_conf->nb_pool_maps; i++) {
fm10k_ethdev.c:1168:16: error: comparison of integers of different signs:
'uint32_t' (aka 'unsigned int') and 'int' [-Werror,-Wsign-compare]
From: Julien Meunier
Once posted through mailbox, we must check for nack from the PF and report an
error in this case.
Signed-off-by: Julien Meunier
Signed-off-by: David Marchand
---
drivers/net/e1000/igb_ethdev.c | 10 +-
1 file changed, 9 insertions(+), 1 deletion(-)
diff --git a/d
From: Julien Meunier
Report an error when something went wrong.
Signed-off-by: Julien Meunier
Signed-off-by: David Marchand
---
drivers/net/enic/enic_ethdev.c | 7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/drivers/net/enic/enic_ethdev.c b/drivers/net/enic/enic_ethd
Hmm, I see. It's good to fix this in the generated e-mails between format-patch
and send-email calls. I always review those to be sure they meet my
expectations ;).
Anyway, it is not clear, what has changed in the v3. Just the rte_cycles? You
should explain that at least in the patch. Bett
Hi Thomas,
> -Original Message-
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> Sent: Friday, October 30, 2015 5:01 PM
> To: Iremonger, Bernard
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v2 1/1] vhost_xen: fix compile error in
> main.c
>
> 2015-10-22 14:21, Be
2015-10-30 16:41, Thomas Monjalon:
> 2015-10-30 14:24, Liang, Cunming:
> > > add vector PMD RX for FVL
> > > add vector PMD TX for FVL
> > > add vector PMD scatter RX for FVL
> > > add RX and TX selection function for FVL
> >
> > Acked-by: Cunming Liang
>
> Cunming, before accepting some
2015-10-30 22:16, Zhe Tao:
> --- a/doc/guides/rel_notes/release_2_2.rst
> +++ b/doc/guides/rel_notes/release_2_2.rst
> @@ -39,6 +39,10 @@ Drivers
>Fixed i40e issue that occurred when a DPDK application didn't initialize
>ports if memory wasn't available on socket 0.
>
> +* **i40e: Add vec
2015-10-30 14:24, Liang, Cunming:
> > add vector PMD RX for FVL
> > add vector PMD TX for FVL
> > add vector PMD scatter RX for FVL
> > add RX and TX selection function for FVL
>
> Acked-by: Cunming Liang
Cunming, before accepting some patches, you should run checkpatch
and check the wor
1 - 100 of 363 matches
Mail list logo
